There is a dark side of the AI revolution that could pose a huge risk to the integrity of academia – students using artificially intelligent ChatGPT to cheat. Experts have estimated that over half of college students are already using the AI-based software to write essays and exams – and the number could rise to 90 percent or higher in the future.
Rehan Haque, of artificial intelligence company Metatalent.ai, warns that AI tools could write entire projects and reword them so that AI is undetectable. He even predicts that at the current rate of growth, more than 90 percent of students could soon be resorting to AI-fueled cheating.
The issue is especially pressing considering OpenAI’s new GPT-4 update, which is able to score 90 percent or more on a huge number of exams, including the American bar exam. This AI is even capable of writing short essays like a bio of a famous person, in a matter of seconds.
While OpenAI is working on a tool to detect AI-written content, it is not 100 percent accurate. This is why some school boards across America, such as the Los Angeles Unified School District, Seattle Public Schools, the New York Department of Education, and Oakland Unified, have already taken action, banning or preventing the use of ChatGPT. New York’s Yeshiva College has even updated their cheating policy to make sure using AI-written content is not allowed.
In response, Haque suggests educators explore other measures such as technological regression, meaning they could consider written tests as an alternative to AI cheating. He believes that banning AI is not completely effective and urges teachers to embrace its potential when used in the classroom. He adds that students are motivated and typically rely on advanced technology as it is seen as a tool to help them succeed.
Research conducted by GoStudent also supports this, with 59 percent of young people wanting to see advanced tech like AI on their curriculum. This can help free educators from rote memorization and regurgitation of large amounts of text.
Narmeen Makhani, executive director of the ETS AI Labs, which make tools that can detect AI cheating, notes that the AI can also be used in a positive manner. It can help learners process and summarize important information faster and more efficiently while in school.
Generative AI detection tools, like the ones offered by her company, look for minute patterns such as the lack of any inconsistencies or the use of repetitive simple words. AI detection models also compare text that appears chaotic with that that comes off as less confusing, which is usually AI-generated.
Makhani insists that educators not only attempt to detect AI-generated content, but also use AI in the classroom. She believes this technology can help to assess learners in more helpful and modern ways. It can also help them to develop more significant skills that are useful in the workplace of tomorrow.
At the end of the day, the risks posed by AI technology like ChatGPT can only be effectively counteracted if both educators and students understand the power of this software, how to use it appropriately, and the methods used to detect it.
